Turbo fan engine mathematical model is a highly complex nonlinear system. Solving engine mathematical model with traditional iteration methods turns out to be difficult as these methods are very sensitive to initial values. Therefore particle swarm optimization is used to solve the model. An improved particle swarm optimization algorithm is produced. The mechanism of immune is introduced in the new algorithm. Clone selection mechanism based on Logistic chaotic mutation and diversity maintaining based on probability have been designed. Results show that the proposed algorithm has better searching performance and convergence speed than other compared algorithms when modeling a mixed exhaust turbofan engine.